Coordination (p. 47) [via PaperHive@docloop]

[docloop[bot]]

Annotation imported from PaperHive.

Regarding this part:

The optionality of the template call in the top figure

Ash Asudeh wrote:

I think the argument from coordination is much better than your previous arguments. In order to do the argument remapping that the benefactive does, it has to apply to the verb 'made', not to the conjunction of 'offered' and 'made'. But I don't think this is ultimately a problem for us. I'll explain why. Note, first, that 'offered' is not a benefactive verb, but is rather a ditransitive: i) She offered me an espresso. ii) She offered an espresso to/*for me. So your mooted solution below (58) is not right, either: It would not do to introduce the benefactive information at the lexical level such that the coordinated verbs have the same selectional requirements. In fact, the benefactive information should just be associated with 'made'. The pronoun 'me' is the RECIPIENT of 'offer', but the BENEFICIARY of 'made'. The coordination here is essentially an example of non-constituent ditransitive coordination of the kind discussed by Dowty in CG. But, in fact, our rule and our glue meaning constructors together do get the right result here. First, as already noted, the two DPs (and even the V) are optional in the rule. And the call to the BENEFACTIVE template is also optional. Thus the first V in the conjunction ('offered') is generated without BENEFACTIVE and without its two DP arguments. The second V ('made') is generated with BENEFACTIVE and without its two DP arguments. This means that in fact this is VP conjunction such that you have [[VP [V offered]] CONJ [VP [V made]]] The normal LFG theory of conjunction then distributes 'me' as the OBJ of both 'offered' and 'made' and 'an espresso' as the OBJ_\theta of both. Is there independent evidence that this is VP conjunction? I think so, because you can have VP adverbs that clearly modify one verb not the other: She offered and reluctantly made me an espresso.

_Link to original comment. About docloop._

[docloop[bot]]

Annotation imported from PaperHive.

Stefan Müller wrote:

I first was very sad since what you write seems to break my argument from coordination. But in fact the solution that you suggest does not involve any NP arguments and hence proves my point. > [[VP [V offered]] CONJ [VP [V made]]] There is nothing left of the original constructional phrasal configuration. Therefore the presence of the NP elements must follow from lexical properties of the verbs. I have to think more about this ...

_Link to original comment. About docloop._

[docloop[bot]]

Annotation imported from PaperHive.

Stefan Müller wrote:

Note also that what you wrote before the adverb example is exactly what we intended with the example: > So your mooted solution below (58) is not right, either: It would not do to introduce the benefactive information at the lexical level such that the coordinated verbs have the same selectional requirements. > In fact, the benefactive information should just be associated with 'made'. The pronoun 'me' is the RECIPIENT of 'offer', but the BENEFICIARY of 'made'. Both calls to the benefactive template in (59) are optional. So one call (the one for the first conjunct = offer) would not apply and the second one would apply making the verb ditransitive (as a lexical rule would do).

_Link to original comment. About docloop._